Create app.py

app.py

@@ -0,0 +1,44 @@
+import streamlit as st
+from transformers import pipeline
+
+@st.cache_resource
+def load_classifier(model_path: str):
+    # Loads the URLGuardian classifier from the Hugging Face Hub.
+    return pipeline("text-classification", model=model_path)
+
+# App Title and description
+st.title("URL Typosquatting Detection with URLGuardian")
+st.markdown(
+    "This app uses the **URLGuardian** classifier by Anvilogic from the Hugging Face Hub to detect potential typosquatting. "
+    "Enter a legitimate URL and a potentially typosquatted URL to see the classifier's prediction."
+)
+
+# Load the classifier model from Hugging Face
+model_path = "Anvilogic/URLGuardian"  # Model repository on Hugging Face
+classifier = load_classifier(model_path)
+
+# URL inputs
+url = st.text_input("Enter the URL:", value="https://example.com")
+
+# Typosquatting detection on button click
+if st.button("Check Typosquatting"):
+    if legit_url and suspect_url:
+        
+        result = classifier(url)[0]
+        label = result['label']
+        score = result['score']
+        
+        # Adjust the label names as per the model's documentation.
+        # This example assumes the label for a typosquatted URL might include "typo".
+        if "typo" in label.lower():
+            st.success(
+                f"The model predicts that '{suspect_url}' is likely a typosquatted version of '{legit_url}' "
+                f"with a confidence of {score * 100:.2f}%."
+            )
+        else:
+            st.warning(
+                f"The model predicts that '{suspect_url}' is NOT likely a typosquatted version of '{legit_url}' "
+                f"with a confidence of {score * 100:.2f}%."
+            )
+    else:
+        st.error("Please enter both a legitimate URL and a potentially typosquatted URL.")